Runs on schedule #130
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: PaddlePaddle Gpu Packages Install Check | |
| run-name: Runs on ${{ github.event_name }} | |
| on: | |
| workflow_dispatch: | |
| workflow_call: | |
| schedule: | |
| - cron: "0 20 * * *" | |
| jobs: | |
| run-check: | |
| runs-on: [self-hosted, GPU-2Card] | |
| strategy: | |
| max-parallel: 2 | |
| matrix: | |
| cuda-version: ['cu118', 'cu126'] | |
| images_version: | |
| - InstallCheck-Py3.10-alalinux8 | |
| - InstallCheck-Py3.10-ubuntu20.04 | |
| - InstallCheck-Py3.10-ubuntu22.04 | |
| - InstallCheck-Py3.10-ubuntu24.04 | |
| steps: | |
| - name: Run Paddle installation check in Docker | |
| id: run_check | |
| continue-on-error: true | |
| run: | | |
| echo "Testing Cuda${{ matrix.cuda-version }} on ${{ matrix.images_version }}" | |
| result=0 | |
| docker run --rm --gpus all --net=host \ | |
| ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddleqa:${{ matrix.images_version }} \ | |
| /bin/bash -ec ' | |
| python3 -m pip install --pre paddlepaddle-gpu -i https://www.paddlepaddle.org.cn/packages/nightly/${{ matrix.cuda-version }}/ && | |
| python3 -c "import paddle; paddle.version.show(); paddle.utils.run_check()" | |
| ' || result=$? | |
| echo "result=$result" | |
| if [ $result -ne 0 ]; then | |
| echo "${{ matrix.cuda-version }} | ${{ matrix.images_version }}" >> failed.txt | |
| fi | |
| exit 0 | |
| - name: Upload individual failure result | |
| if: always() | |
| uses: actions/upload-artifact@v4 | |
| with: | |
| name: failed-${{ matrix.cuda-version }}-${{ matrix.images_version }} | |
| path: failed.txt | |
| if-no-files-found: ignore | |
| collect-failures: | |
| runs-on: ubuntu-latest | |
| needs: run-check | |
| steps: | |
| - name: Download all failure artifacts | |
| uses: actions/download-artifact@v4 | |
| with: | |
| path: artifacts | |
| - name: Merge all failed.txt files (deduplicated) | |
| run: | | |
| mkdir -p merged | |
| find artifacts -name 'failed.txt' -exec cat {} + | sort | uniq > merged/all_failed.txt || true | |
| - name: Upload merged failure list | |
| uses: actions/upload-artifact@v4 | |
| with: | |
| name: failed-combos | |
| path: merged/all_failed.txt | |
| if-no-files-found: ignore | |
| create-issue-if-needed: | |
| runs-on: ubuntu-latest | |
| needs: collect-failures | |
| if: always() | |
| steps: | |
| - name: Download merged failure list | |
| uses: actions/download-artifact@v4 | |
| with: | |
| name: failed-combos | |
| path: artifacts | |
| - name: Create issue if there are failures | |
| uses: actions/github-script@v7 | |
| with: | |
| github-token: ${{ secrets.GITHUB_TOKEN }} | |
| script: | | |
| const fs = require('fs'); | |
| const path = 'artifacts/all_failed.txt'; | |
| const issueTitle = '❌ Paddle GPU 安装验证失败(Nightly)'; | |
| if (!fs.existsSync(path)) { | |
| console.log('没有失败记录，无需创建 issue'); | |
| return; | |
| } | |
| const failedVersions = fs.readFileSync(path, 'utf-8').split('\n').filter(Boolean); | |
| if (failedVersions.length === 0) { | |
| console.log('all_failed.txt 文件为空'); | |
| return; | |
| } | |
| const existingIssues = await github.rest.issues.listForRepo({ | |
| owner: context.repo.owner, | |
| repo: context.repo.repo, | |
| state: 'open', | |
| labels: 'auto-generated', | |
| per_page: 100 | |
| }); | |
| const alreadyExists = existingIssues.data.some(issue => | |
| issue.title.includes(issueTitle) | |
| ); | |
| if (alreadyExists) { | |
| console.log('已有相同标题的 open issue，跳过创建'); | |
| return; | |
| } | |
| const list = failedVersions.map(v => `- ❌ ${v.trim()}`).join('\n'); | |
| const body = ` | |
| 以下 CUDA + 系统镜像组合的 Paddle 安装检查失败： | |
| ${list} | |
| Workflow: [查看运行详情](${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}) | |
| @XieYunshen 请关注此问题。 | |
| `; | |
| await github.rest.issues.create({ | |
| owner: context.repo.owner, | |
| repo: context.repo.repo, | |
| title: issueTitle, | |
| body, | |
| labels: ['ci-failure', 'gpu', 'auto-generated'], | |
| assignees: ['XieYunshen'] | |
| }); |